Nano Banana Pro CN

图片生成技能，当用户需要生成图片、视觉信息图、创建图像、编辑/修改/调整已有图片时使用此技能。基于中国的API易代理站(https://apiyi.com/)的NanoBananaPro模型的图片生成服务，无需访问外网。支持10种宽高比的图片比例（`1:1`、`16:9`、`9:16`、`4:3`、`3:4`、`...

This skill's code matches its advertised image-generation purpose, but be cautious before installing: - Secrets & config mismatch: the scripts expect APIYI_API_KEY, SKILL.md says the same, but skill.yaml names the key API_KEY and the registry metadata claims no required env vars. Fix or confirm which name your platform will store the secret under to avoid leaking or failing to provide the key. - Privacy: input images are base64-encoded and uploaded to api.apiyi.com. Do not use this skill with sensitive or private images unless you trust that third party and the account you register. - Dependencies: Node version uses built-in modules (no extra deps); the Python script requires the 'requests' package but no dependency is declared—ensure 'requests' is installed in the runtime environment. - Provenance: source/homepage metadata is inconsistent (skill.yaml lists a homepage, top-level metadata said none and source is unknown). Verify the publisher/owner before trusting or sharing credentials. Recommended immediate actions: inspect and (if possible) correct the skill.yaml to declare the correct secret name (or map API_KEY → APIYI_API_KEY), install and test the Python/Node runtime in an isolated environment, and avoid uploading sensitive input until you're satisfied with the endpoint and account settings. If you cannot confirm the publisher or the correct configuration, treat this skill as untrusted.

Type: OpenClaw Skill Name: nano-banana-pro-cn Version: 1.0.0 The skill bundle provides image generation and editing functionality via the NanoBananaPro model using a documented Chinese proxy service (apiyi.com). The implementation includes both Node.js (generate_image.js) and Python (generate_image.py) scripts that handle standard API interactions, base64 image encoding, and local file saving. No evidence of malicious behavior, data exfiltration, or harmful prompt injection was found; the instructions in SKILL.md are appropriately focused on parameter extraction and task execution.